Three AI Assistants Move Into Your Browser

In early 2026, all three major AI companies quietly moved their assistants into the browser. Google was first: Gemini appeared as a persistent sidebar inside Chrome in January, sitting on the right side of the screen no matter which tab you were on. From there it can read the page you are viewing, answer questions about it, and connect to Gmail, Calendar, and Google Maps without you switching apps. For users on AI Pro or Ultra plans, an "auto browse" mode lets Gemini carry out multi-step tasks on its own, such as comparing prices across several websites.

Anthropic's Claude in Chrome arrived in beta shortly after, available to Max, Team, and Enterprise subscribers. Instead of a sidebar, it works through a chat interface that can control the browser directly. Users can drag tabs into a shared group so Claude can read and interact with all of them at once. It also supports scheduled tasks — you can tell it to run a routine daily or weekly — and it integrates with outside services like Uber and Booking.com. Developers get an extra advantage: Claude can read the browser console output, including network errors, without leaving the browser.

OpenAI's Codex Chrome extension launched on 8 May 2026, taking a different approach altogether. Rather than acting as a sidebar or a chat window you open yourself, Codex operates in background tabs while you keep using the browser normally. It was designed for workplace tools: it can log into LinkedIn, Salesforce, or Gmail using your existing signed-in session and then fill forms, update records, or move through multi-step workflows across those apps. Privacy controls are built in — Codex asks before it accesses any new website or your browser history, and OpenAI says it only stores data that becomes part of the active conversation thread. All three tools show that the AI assistant is no longer a separate app you visit; it is becoming a layer that lives inside the browser itself.